The Contribution of Hepatic Inactivation of Testosterone to the Lowering of Serum Testosterone Levels by Ketoconazole

TOXICOLOGICAL SCIENCES 54, 128 137 (2000) Copyright 2000 by the Society of Toxicology The Contribution of Hepatic Inactivation of Testosterone to the Lowering of Serum Testosterone Levels by Ketoconazole Vickie S. Wilson and Gerald A. LeBlanc 1 Department of Toxicology, North Carolina State University, Raleigh, North Carolina 27695 Received July 9, 1999; accepted November 1, 1999 Hepatic biotransformation processes can be modulated by chemical exposure and these alterations can impact the biotransformation of endogenous substrates. Furthermore, chemically mediated alterations in the biotransformation of endogenous steroid hormones have been implicated as a mechanism by which steroid hormone homeostasis can be disrupted. The fungicide ketoconazole has been shown to lower serum testosterone levels and alter both gonadal synthesis and hepatic inactivation of testosterone. The present study examined whether the effects of ketoconazole on the hepatic biotransformation of testosterone contribute to its lowering of serum testosterone levels. Results also were used to validate further the use of the androgen-regulated hepatic testosterone 6 /15 -hydroxylase ratio as an indicator of androgen status. Male CD-1 mice were fed from 0 to 160 mg/kg ketoconazole in honey. Four h after the initial treatment, serum testosterone levels, gonadal testosterone secretion, and hepatic testosterone hydroxylase activity decreased, and the hepatic testosterone 6 / 15 -hydroxylase ratio increased in a dose-dependent manner. Immunoblot analysis indicated that the transient decline in hepatic biotransformation was not due to reduced P450 protein levels. Rather, hepatic testosterone biotransformation activities were found to be differentially susceptible to direct inhibition by ketoconazole. Differential inhibition was also responsible for the increase seen in the 6 /15 -hydroxylase ratio. The changes in serum testosterone levels could be explained by decreased gonadal synthesis of testosterone and were not impacted by decreased hepatic biotransformation of testosterone. These results demonstrate that changes in the hepatic hydroxylation of testosterone by ketoconazole, and perhaps other chemicals, have little or no influence serum testosterone levels. Key Words: androgen disruption; hepatic biotransformation; ketoconazole; mice; testosterone. INTRODUCTION 1 To whom correspondence should be addressed at Department of Toxicology, North Carolina State University, Box 7633, Raleigh, NC 27695. Fax: (919) 515-7169. E-mail: gal@unity.ncsu.edu. Testosterone is metabolically inactivated in the liver by some of the same hepatic enzymes responsible for the detoxification of xenobiotics. It is well established that the expression of these hepatic enzymes is differentially susceptible to induction or suppression due to xenobiotic exposure (Gonzalez, 1989; Guengerich and Shimada, 1991; Kulkarni and Hodgson, 1984). As a result, alterations in hepatic enzyme activity due to chemical exposure can lead to concomitant changes in the biotransformation of endogenous substrates such as testosterone. It has been further suggested that changes in the rate of inactivation and elimination of testosterone, as a consequence of chemical exposure, can alter the levels of circulating testosterone and thereby disrupt steroid hormone homeostasis (Bammel et al., 1992; den Besten et al., 1991; Gagnon et al., 1994; Gradowska-Olszewska et al., 1984; Haake et al., 1987; Machala et al., 1998; McMaster et al., 1991; Sanderson et al., 1997; Sivarajah et al., 1978; Tredger et al., 1984). In previous studies, we have demonstrated that chemical exposure can cause increased hepatic biotransformation and elimination of testosterone (Wilson and LeBlanc, 1998). However, these changes in the metabolic elimination of testosterone did not significantly impact serum testosterone levels, presumably due to adequate regulatory feedback mechanisms. Nevertheless, chemical effects on the hepatic biotransformation of testosterone may contribute to perturbations in serum androgen levels when testosterone synthesis (i.e., feedback compensation) is compromised. Whether or not chemically mediated changes in hepatic biotransformation of testosterone can impact androgen homeostasis is an important issue. Steroid hormones play decisive and irreversible roles in embryonic development and sex differentiation (vom Saal et al., 1992). In addition, they influence the acquisition and maintenance of secondary sex characteristics in adults (Grumbach and Conte, 1981) and are critical in ensuring adult male and female reproductive function (Hadley, 1982). Abnormalities in these processes generally occur when there are changes in the amount or availability of a particular hormone or due to differences in the hormone s pattern of secretion (Wilson and Foster, 1985). Therefore, chemically induced modulations in hepatic biotransformation capabilities (either inductive or suppressive) that impact circulating steroid hormone levels have the potential to produce dramatic effects on steroid hormone regulated processes. Circulating testosterone levels are maintained by a dynamic 128

IMPACT OF METABOLISM ON SERUM ANDROGEN 129 system that ensures androgen homeostasis. In simplest terms, this system can be considered a balancing act between the rate of testosterone synthesis and the rate of its metabolic inactivation/elimination. If this system is truly a balancing act, then both the rate of synthesis and the rate of metabolic inactivation/ elimination impact the maintenance of circulating testosterone levels, and chemically induced disruptions in either process can potentially modify circulating testosterone levels. In the present study, ketoconazole was used as a model compound to further investigate the role of the metabolic inactivation of testosterone on the maintenance of circulating androgen levels. Our goal was to determine whether the modulation of the hepatic biotransformation of testosterone by ketoconazole, in combination with its effects on testosterone synthesis, would contribute to the lowering of serum testosterone levels by this compound. These studies were also used to continue evaluation our recent observation that the ratio of hepatic testosterone 6 /15 -hydroxylase activities provides a superior indicator of the androgen status in CD-1 mice (Wilson et al., 1999). These two hydroxylase activities are differentially regulated by testosterone such that a high ratio is typical of females or indicative of demasculinization, and a low ratio is typical of males or indicative of masculinization. Ketoconazole has been shown to significantly lower serum testosterone levels (Pont et al., 1982). If so, then it should also dramatically affect the testosterone 6 /15 -hydroxylase ratio. The lowering of serum testosterone levels by ketoconazole treatment has been attributed primarily to the ability of ketoconazole to inhibit gonadal synthesis of testosterone (Pont et al., 1982). There have been, however, conflicting observations regarding the effects of ketoconazole on hepatic metabolic enzymes, with induction, inhibition, or no effect all being reported by different researchers (Jiritano et al., 1986; Morita et al., 1988; Niemegeers et al., 1981; Rodrigues et al., 1988; Ronis et al., 1994; Thomson et al., 1988). In CD-1 mice, testosterone is inactivated/metabolized in the liver by four primary pathways. First, cytochrome P450 isozymes can monohydroxylate testosterone in a characteristic manner that is both region specific and stereospecific (Waxman et al., 1983). Second, testosterone can be dehydrogenated to produce androstenedione through P450-mediated pathways (Waxman, 1988), and also through the action of 17 -hydroxysteroid dehydrogenase (Bloomquist et al., 1977). Although the production of androstenedione can contribute to the androgen precursor pool in the body, we focused primarily on its role as an inactivation product. Third and fourth, UDP-glucuronosyltransferase and sulfotransferase can conjugate testosterone directly or subsequent to hydroxylation to either glucuronic acid or sulfate, respectively. These pathways generally produce more polar products that can then be eliminated from the body. The primary goals of this study were the following: a) assess the temporal effect of ketoconazole on serum testosterone levels; b) establish the relative contribution of gonadal synthesis and hepatic biotransformation of testosterone in altering serum testosterone levels following ketoconazole treatment; and c) evaluate the effects of ketoconazole on the testosterone 6 /15 -hydroxylase ratio, an indicator of androgen status. MATERIALS AND METHODS Animals and treatments. Eight-week old male CD-1 mice were purchased from Charles River Laboratories (Raleigh, NC) and held at our animal facility for at least 1 week prior to initiation of treatment. Animals had free access to food and water throughout the study and were held under controlled temperature (18 21 C) and light (12-h diurnal light cycle) conditions. During treatment, each mouse was fed once a day with ketoconazole suspended in 0.1 ml honey at the concentration required to result in dosages of either 20, 40, 80, or 160 mg ketoconazole per kg body weight per day. Control animals each received 0.1 ml of honey alone. The mice willingly ate aliquots of honey from the blunt end of a 1-cc graduated syringe. Initially, three or four animals from each treatment were euthanized 4 h after receiving a single dose of ketoconazole, 24 h after a single dose, and 24 h after the final dose of seven daily treatments. An additional six to nine animals from each of the control and two highest treatment levels were sacrificed 4 h after a single dose of ketoconazole. All animals were euthanized by CO 2 asphyxiation. Serum testosterone. Blood from each animal was collected by cardiac puncture and allowed to clot at room temperature for 1 h; serum was obtained by centrifugation at 14,000 g for 10 min (LeBlanc and Waxman, 1988). Serum from each sample was then immediately frozen at 20 C until assayed. Total testosterone was measured in prepared serum by solid-phase radioimmunoassay using commercially available reagents and protocols (Diagnostic Products Corp., Los Angeles, CA). Hepatic microsome and cytosol preparation. Livers were excised, weighed, minced in ice-cold buffer, and rapidly frozen in liquid nitrogen. Individual livers were thawed and homogenized on ice in chilled buffer (0.1 M HEPES, ph 7.4, 1mM EDTA, and 10% glycerol) using a glass tissue homogenizer. Microsomes and cytosol were prepared by differential centrifugation (van der Hoeven and Coon, 1974). The cytosolic supernatant was reserved for sulfotransferase assays and the microsomal pellet was resuspended in microsome buffer (0.1 M potassium phosphate, ph 7.4, 0.1 mm EDTA, and 20% glycerol). Protein concentrations were determined (Bradford, 1976) using commercially prepared reagent (Bio-Rad, Hercules, CA.) and bovine serum albumin (Sigma, St. Louis, MO) as standards. Both microsomes and cytosol were stored at 80 C until assays were performed. Testosterone hydroxylase activity. Both the rate of testosterone hydroxylation and the rate of androstenedione production were measured in this assay. These activities were assayed using 400 g microsomal protein and 40 nmol [ 14 C]testosterone as a substrate (1.8 mci/mmol, Dupont NEN, Boston, MA) in 0.1 M potassium phosphate buffer (ph 7.4) as previously described (Baldwin and LeBlanc, 1992). Reactions were initiated by the addition of 1 mm NADPH and incubated for 10 min at 37 C. Product formation was shown to be linear over this time period. The total assay volume was 400 l. After 10 min, the reactions were terminated by the addition of 1 ml ethyl acetate and vortexing. Metabolites were removed from the reaction mixture by differential solvent extraction. The aqueous phase was extracted with ethyl acetate a total of three times (1 ml, 2 ml, 1 ml). After each addition of ethyl acetate, the tubes were vortexed for 1 min, then centrifuged for 10 min to separate the ethyl acetate and aqueous phases. Ethyl acetate fractions derived from each aqueous sample were combined and then dried under a stream of nitrogen. The residues were suspended in 79 l (35 l 2) ethyl acetate and separated by thin-layer chromatography. Individual metabolites and unmetabolized [ 14 C]testosterone were identified as we have previously described (Baldwin and LeBlanc, 1992) and quantified using InstantImager electronic autoradiography (Packard Instrument Co., Meriden, CT). The metabolite identified in the text and figures as 15 /7 -hydroxytestosterone comigrated with both standards during TLC and the precise identity of this metabolite could not be determined. Specific activity

130 WILSON AND LEBLANC for the production of each metabolite was calculated as picomoles of metabolite produced per minute of the assay per milligram of microsomal protein. UDP-glucuronosyltranferase activity. UDP-glucuronosyltransferase activity towards [ 14 C]testosterone was assayed under conditions previously described (Tedford et al., 1991). In brief, microsomal protein (200 g) was incubated at 37 C with 40 nmol [ 14 C]testosterone (1.8 mci/mmol) in 0.1 M potassium phosphate buffer, ph 7.4. Reactions were initiated with 10 l uridine 5 -diphosoglucuronic acid in buffer (12.9 mg/ml, Sigma) to give a total assay volume of 400 l. The reactions were terminated after 10 min by the addition of 2 ml ethyl acetate and vortexing. Product formation was shown to be linear over this time period. Each sample was vortexed for 1 min and aqueous and ethyl acetate phases were separated by centrifugation. The ethyl acetate fraction containing the unconjugated [ 14 C]testosterone was removed and extraction of the aqueous phase was repeated a second time with an additional 2 ml ethyl acetate. [ 14 C]Testosterone-glucuronide was quantified by liquid scintillation counting of a 100- l aliquot of the postextraction aqueous phase. Samples were run with every assay that consisted of all constituents except microsomes. These samples were used to account for any spontaneous conjugation of glucuronic acid to testosterone and for any [ 14 C]testosterone that was not extracted from the aqueous phase. Radioactivity in these aqueous samples following ethyl acetate extraction (typically 0.1% of the total activity in the assay) was subtracted from the total radioactivity associated with the glucuronic acid conjugated testosterone in each assay. Specific activity was calculated as picomoles of conjugate produced per minute per milligram of microsomal protein. Sulfotransferase activity. Cytosolic protein (200 g) was incubated with 40 nmol [ 14 C]testosterone (1.8 mci/mmol) in 0.1 M potassium phosphate buffer (ph 6.5) to assay for the activity of sulfotransferase enzymes toward testosterone. Reactions were initiated with 10 l adenosine 5 -phosphosulfate (10.1 mg/ml) for a total assay volume of 400 l. Assay tubes were covered and incubated in a 37 C water bath for 20 h. Product formation was linear over this time period. Reactions were terminated by addition of 2 ml ethyl acetate and vortexing. Unconjugated testosterone was removed by ethyl acetate extraction (two times, 2 ml each). Sulfate-conjugated [ 14 C]testosterone was identified and quantified as we have previously described (Wilson and LeBlanc, 1998). Specific activity was calculated as picomoles of metabolite produced per minute per milligram of cytosolic protein. Testosterone hydroxylase inhibition assay. Testosterone hydroxylase activities were measured as described above except that a series of concentrations of ketoconazole were added to the reaction mixture. A5mMstock solution of ketoconazole in 100% ethanol was prepared. Working solutions (100 ) for each ketoconazole concentration were prepared by diluting this stock solution in ethanol so that 4 l of ethanol was added to each 400- l assay. Concentrations of ketoconazole in individual assays were 0 (vehicle only), 0.010, 0.030, 0.10, 0.30, 1.0, 3.0, 10, and 30 M. Assays were run at each ketoconazole concentration in triplicate. Testicular synthesis of testosterone. Four h after treatment, the effects of ketoconazole on testicular synthesis of testosterone were evaluated ex vivo with and without human chorionic gonadotropin (hcg, Sigma) stimulation. Both testes were removed from each animal in the control and 80 and 160 mg/kg/day treatment groups. Each testis was immediately transferred into 5 ml of phenol red free Dulbecco s Modified Eagle Medium (Life Technologies, Gaithersburg, MD) supplemented with 10% charcoal-dextran treated fetal bovine serum (Hyclone, Logan, UT). The right testis from each animal was incubated in medium without hcg. The left testis was incubated in medium with 100 miu hcg per ml of medium. Testes were incubated in an atmosphere of 95% air/5% CO 2 for exactly 4 h. Testosterone secreted into the medium was measured by RIA as described for serum testosterone. Preliminary experiments had demonstrated that the testicular secretion of testosterone under these conditions was linear for at least 22 h and that there were no detectable levels of testosterone in medium incubated without a testis. Immunoblotting procedure. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of 10 g microsomal protein was performed using a 10% separating gel with a 4% stacking gel. Dexamethasone induced rat hepatic microsomal protein (Oxford Biomedical Research, Oxford, MI) was included in one lane of each gel as a positive control. Proteins were then electrophoretically transferred onto nitrocellulose at 120 V for 1hincold buffer (25 mm Tris base, 192 mm glycine, 20% v/v methanol, 0.015% w/v sodium dodecyl sulfate). The nitrocellulose was blocked overnight at 4 C in phosphate-buffered saline containing 0.3% Tween 20 (Sigma). After rinsing, the nitrocellulose membrane was incubated with monoclonal rat anti-cyp3a1 antibody (Oxford Biomedical Research) at a dilution of 1:1000 for 2hatroom temperature. The nitrocellulose was again rinsed and then incubated with anti-mouse IgG secondary antibody conjugated to alkaline phosphatase for 2 h at a 1:1000 dilution. Immunoblotted protein was visualized colorimetrically using 5-bromo-4-chloro-indolyl phosphate/nitro blue tetrazolium (Waxman et al., 1988). Relative band intensities were determined using a scanning laser densitometer (Zeineh, Fullerton, CA). Statistics. Statistical significance was determined using ANOVA and Dunnett s Multiple Comparison Test. These statistical analyses were performed using JMP statistical software (SAS Institute, Cary, NC). IC50 calculations and linear regression analyses were performed using Origins statistical/ graphing software (Microcal, Northhampton, MA). RESULTS All of the mice used in this study survived chemical treatment with no signs of physical distress or loss of appetite for food or water. The animals in the 7-day treatment group gained weight comparable to controls during the treatment period. Liver weights and liver/body weight ratios were not significantly affected at any treatment level. To assess the temporal effects of ketoconazole on serum testosterone levels, initial studies were conducted utilizing serial sacrifice of groups of mice. Serum testosterone and testosterone biotransformation were evaluated 4 h after the initial dosage, 24 h after the initial dosage, and after 7 days treatment with 20, 40, or 80 mg/kg/ day ketoconazole. Four hours after treatment with ketoconazole, serum testosterone levels decreased in a dose-dependent manner (Fig. 1A). Twenty-four hours after the initial treatment and after seven daily treatments the changes in serum testosterone levels were variable and did not exhibit the same pattern of effects as seen after 4 h (Figs. 1B and 1C). None of the changes in serum testosterone levels were statistically significant as compared to controls, due to the high level of variability within treatments. We next examined the effects of ketoconazole on testosterone biotransformation processes at all three time points to determine if the lowering of serum testosterone levels correlated with any increases in hepatic biotransformation. Only P450-mediated hydroxylation was affected by ketoconazole treatment, and hydroxylation was decreased, not increased (Table1). Ketoconazole treatment resulted in a statistically significant dose-dependent decrease in the rate of hydroxytestosterone metabolite formation 4 h after dosing. At the 20, 40, and 80 mg/kg/day dosages, total testosterone hydroxylase activity was decreased by 42%, 48%, and 53%, respectively. There were no significant effects on oxidoreduction, glucuronic acid conjugation, or sulfate conjugation of testosterone due to ketoconazole treatment (Table 1). Twenty-four hours

IMPACT OF METABOLISM ON SERUM ANDROGEN 131 FIG. 1. Effects of ketoconazole on serum testosterone levels after 4 h, 24 h, and 7 days treatment. Serum testosterone levels in male mice treated with 0, 20, 40,or 80 mg/kg/day ketoconazole (A) 4 h after a single dosage, (B) 24 h after a single dosage, and (C) after seven daily treatments. Mean and standard deviation (n 3 4) is presented as a percent of the serum testosterone levels obtained in controls. after a single dose of ketoconazole, all testosterone biotransformation activities were comparable to those seen in control animals (Table1). These parameters were also comparable to controls after 7 days of treatment. Because serum testosterone levels decreased in a dose-dependent manner and testosterone hydroxylase activities were significantly affected only at the 4-h time point, additional studies to examine the contribution of testosterone biotransformation to the maintenance of circulating testosterone levels were conducted using this time point. As we have previously shown (Baldwin and LeBlanc, 1992), total testosterone hydroxylation represents the combined conversion of testosterone to at least seven different chromatographically distinguishable monohydroxytestosterone metabolites. Also, we have previously demonstrated that the ratio of testosterone 6 -hydroxylase and 15 -hydroxylase activities can serve as a biomarker of the androgen status in CD-1 mice without the high interindividual variability typically found in serum testosterone measurements in mice (Wilson et al., 1999). The effect of ketoconazole treatment on the rate of production of individual hydroxytestosterone metabolites was evaluated 4 h after dosing (Fig. 2). The decrease in total testosterone hydroxylase activity was due to significant dosedependent decreases in the rates of production of three metabolites. Testosterone 6 -, 15 -, and 2 -hydroxylase specific activity were differentially suppressed due to ketoconazole treatment. 6 -Hydroxytestosterone metabolite production was the most dramatically reduced. The testosterone 6 /15 -hydroxylase (6 /15 -OH) ratios were calculated for the ketoconazole treated mice (Fig. 3). Four hours after ketoconazole treatment, the 6 /15 -OH ratio increased as serum testosterone levels decreased, which is indicative of demasculinization. However, the increase in the ratio was not statistically significant and the variability associated with the ratio was higher than expected based on earlier studies using other compounds (Wilson et al., 1999). The observed changes in the ratio reflected a decrease in the rate of production of 15 -hydroxytestosterone, but no effect on the production of 6 -hydroxytestosterone. These observations in conjunction with the rapid suppression and recovery of hydroxylase activities suggest that the effects of ketoconazole may have been due to direct inhibition and not changes in the level of protein expression. To determine if the decrease in testosterone hydroxylase activity was due to reduced enzymatic protein levels, immunoblots were performed using hepatic microsomal protein from the ketoconazole-treated mice. As testosterone 6 -hydroxylase activity was most dramatically affected by ketoconazole treatment, the level of microsomal CYP 3A, the protein responsible for this activity, was determined using a monoclonal antibody to rat P450 CYP3A1 protein. This antibody was shown to be highly specific for mouse CYP3A protein. One immunoreactive protein was detected in each hepatic microsomal sample from treated or control mice. Figure 4A depicts a representative immunoblot comparing the levels of CYP 3A protein in hepatic microsomes from the ketoconazole-treated and control mice. No significant difference in staining intensity was detected between control and ketoconazole-treated animals. Because reduced protein levels were not responsible for the decreased activity observed in testosterone 6 -hydroxylase activity, ketoconazole was evaluated for its ability to directly inhibit testosterone hydroxylase activities. In vitro incubation of pooled hepatic microsomes from untreated male CD-1 mice with increasing concentrations of ketoconazole demonstrated that several testosterone hydroxylase activities were directly inhibited by ketoconazole. Testos-

132 WILSON AND LEBLANC TABLE 1 Hepatic Testosterone Biotransformation Specific Activities After Ketoconazole Treatment Dosage (mg/kg) Product Control 20 40 80 4 h after treatment Total hydroxytestosterone 6002 1061 3482 716* 3102 403* 2807 1002* Androstenedione 980 112 844 212 996 134 842 314 Glucuronic acid conjugate 68.0 6.83 85.2 23.9 59.5 9.26 87.9 7.79 Sulfate conjugate 0.202 0.076 0.195 0.10 0.160 0.012 0.238 0.033 24 h after treatment Total hydroxytestosterone 7874 467 6531 1197 7336 1319 6893 2273 Androstenedione 806 78.2 847 173 684 80.7 786 131 Glucuronic acid conjugate 90.6 3.37 73.8 17.9 83.39 18.2 93.03 18.4 Sulfate conjugate 0.232 0.125 0.166 0.149 0.219 0.103 0.345 0.112 After 7 days treatment Total hydroxytestosterone 7626 2568 8787 1126 7931 184 7565 584 Androstenedione 1195 123 932 168 1045 57.9 1060 81.6 Glucuronic acid conjugate 78.2 5.39 80.5 11.3 76.0 9.33 98.3 14.9 Sulfate conjugate 0.193 0.124 0.209 0.070 0.243 0.065 0.253 0.045 Note. Data are presented as mean specific activity (pmol/min/mg protein) standard deviation (n 3 4). * Indicates a statistically significant difference when compared to controls (p 0.01). terone 6 -hydroxylase activity was most dramatically affected by the addition of increasing concentrations of ketoconazole in these assays (Fig. 4B). The IC50 for ketoconazole inhibition of testosterone 6 -hydroxylase activity was 0.109 0.015 M. Four testosterone hydroxylase activities, 6, 15, 15 /7, and 16, were affected by ketoconazole, albeit to different degrees. The IC50 values for each testosterone hydroxylase activity are shown in Table 2. These results, combined with the immunoblot data, indicated that the decrease in testosterone hydroxylase activity could be attributed to direct inhibition by ketoconazole. Concentrations of ketoconazole over 30 M were not evaluated because ketoconazole could not be reliably maintained in solution at higher concentrations. The production of 6 -hydroxytestosterone in vitro was not appreciably affected by ketoconazole (Fig. 5A). The production of 15 -hydroxytestosterone, however, was inhibited by ketoconazole (Fig. 5B). Increases in the 6 /15 -OH ratio, therefore, resulted from the FIG. 2. Effect on individual hepatic testosterone hydroxylase activities 4 h after treatment with ketoconazole. Each set of four bars illustrates the effect of ketoconazole treatment on the production of an individual hydroxytestosterone metabolite. Data are presented as mean and standard deviation (n 3 4). All treatments significantly (p 0.005) reduced 2, 6, and 15 hydroxylase activities. FIG. 3. The testosterone 6 :15 -hydroxylase ratio 4 h after ketoconazole treatment. Data are represented as mean and standard deviation (n 3 4).

IMPACT OF METABOLISM ON SERUM ANDROGEN 133 FIG. 4. (A) Western blot of hepatic microsomal cytochrome P450 CYP3A protein from mice exposed to ketoconazole for 4 h. (B) Direct inhibition of testosterone 6 -hydroxylase specific activity by ketoconazole. Each data point in the in vitro testosterone 6 -hydroxylase activity inhibition curve is the mean of three individual assays at each ketoconazole concentration. IC50 value was calculated from a sigmoid curve fitted to the inhibitor-response data using Origin data analysis software (MicroCal Software, Northhampton, MA). differential inhibition of 15 -hydroxylase by ketoconazole and were not due to the differential regulation of the production of these metabolites by androgen. Results indicated that ketoconazole inhibited the hepatic biotransformation of testosterone. This effect would not contribute to the observed decrease in serum testosterone levels. In fact, the inhibition of hepatic testosterone biotransformation may have attenuated any decrease in serum testosterone levels caused by ketoconazole. Experiments were next conducted to determine whether the decrease in serum testosterone levels could be explained exclusively by the effects of ketoconazole on gonadal testosterone secretion or whether the resultant serum testosterone levels could best be attributed to the combined effects of ketoconazole on both gonadal testosterone secretion and hepatic testosterone biotransformation. Mice were dosed with 0, 80, or 160 mg/kg ketoconazole and the effects on serum testosterone levels, gonadal testosterone secretion, and hepatic testosterone hydroxylation were evaluated 4 h later. Serum testosterone levels, gonadal testosterone secretion, and hepatic testosterone hydroxylation were all significantly reduced in both the 80 and 160 mg/kg ketoconazole treatment groups (Fig. 6). hcg stimulation increased overall testosterone secretion by isolated testes, but did not overcome the inhibitory effect of ketoconazole on testicular testosterone secretion. With hcg stimulation, testicular secretion of testosterone was decreased by 27% and 49% of control levels in the 80 and 160 mg/kg ketoconazole treatment groups, respectively. Comparison of serum testosterone levels and gonadal testosterone secretion among individual mice revealed that these parameters were highly correlated (r 0.963, p 0.001; Fig. 7A). Similar comparison between serum testosterone levels and hepatic testosterone hydroxylation revealed that ketoconazole-treated mice exhibited both low serum testosterone levels and low hepatic biotransformation capabilities, whereas control mice exhibited high hepatic hydroxylation capability and variable serum testosterone levels (Fig. 7B). If control and treated animals are evaluated separately, then the regression lines in both cases are virtually horizontal and there is no relationship between the two parameters in either case (controls: r 0.14, p 0.72; treated: r 0.03, p 0.92). When data for both controls and treated are combined, a positive relationship is observed (r 0.59, p 0.003), although of lesser significance than that observed with serum testosterone levels and gonadal testosterone secretion. Finally, serum testosterone levels were compared to the ratio of the rates of gonadal secretion of testosterone and hepatic hydroxylation of testosterone to determine whether the combination of these two parameters might best explain serum testosterone levels (Fig. TABLE 2 Concentrations of Ketoconazole That Resulted in Fifty Percent Inhibition (IC50) of Individual Testosterone Hydroxylase Activities Hydroxytestosterone metabolite IC50, M Error, M 6 0.109 0.02 15 0.239 0.03 15 /7 2.26 1.02 16 25.4 2.24 2 30.0 N/A 6 30.0 N/A 16 30.0 N/A

134 WILSON AND LEBLANC FIG. 5. Dose-response curves showing the effect of ketoconazole on the specific activity of (A) testosterone 6 -hydroxylase and (B) testosterone 15 -hydroxylase. Each data point is the mean of assays run in triplicate at each concentration of ketoconazole. 7C). The relationship between this ratio and serum testosterone levels (r 0.904, p 0.001) was comparable to that observed between serum testosterone levels and gonadal secretion, indicating that hepatic biotransformation did not significantly impact serum testosterone levels. DISCUSSION Our results indicate that ketoconazole s ability to lower serum testosterone levels in CD-1 mice is transient, necessitating short-term (4-h) exposures in order to evaluate our hypothesis. These results agree with the marked but transient drop in serum testosterone levels demonstrated in humans receiving ketoconazole therapy (Pont et al., 1982; Santen et al., 1983) and in rats and mice treated with ketoconazole (Heckman et al., 1992). In addition, the only effect on hepatic biotransformation processes that we observed was the transient decrease of some testosterone hydroxylase activities. Other testosterone biotransformation processes (i.e., oxidoreduction, conjugation) were not affected by ketoconazole treatment. The mechanism responsible for the transient decrease in some hepatic hydroxylase activities following ketoconazole treatment became apparent during the in vitro assessment of the effects of ketoconazole on hepatic testosterone hydroxylase activities. Following in vivo treatment of mice with ketoconazole, 6 - and 15 -hydroxylase activities were significantly decreased and 15 /7 -hydroxylase activity was slightly decreased. Testosterone 16 -, 6 -, and 16 -hydroxylase activities were not affected by treatment with ketoconazole. In vitro experiments with liver microsomal preparations demonstrated that testosterone 6 - and 15 -hydroxylase activities were highly susceptible to direct inhibition by ketoconazole, 15 / 7 -hydroxylase activity was moderately susceptible to inhibition, and 16 -, 6 -, and 16 -hydroxylase activities were rather insensitive to inhibition by ketoconazole. Thus, the transient decrease in some of the testosterone hydroxylase activities may be attributed to direct inhibition by ketoconazole. Hepatic testosterone 2 -hydroxylase activity was also transiently reduced following treatment of animals with ketoconazole. However, this enzymatic activity was not inhibited during in vitro incubations with ketoconazole. The rapid and transient nature of this reduction in activity after in vivo treatment suggests that this enzyme also was inhibited. These observations raise the possibility that a metabolite of ketoconazole generated in vivo was responsible for the inhibition of this enzyme. Ketoconazole treatment resulted in an increase in the testosterone 6 /15 -hydroxylase ratio. Normally, an increase in this ratio is indicative of demasculinization because of the negative effect of testosterone on 6 -hydroxylase activity and its positive effect on 15 -hydroxylase activity (Wilson et al., 1999). Although the increase in the ratio was coincidental with the decrease in serum testosterone levels following ketoconazole treatment, altered serum testosterone levels did not mediate this effect. Rather, ketoconazole differentially inhibited testosterone 15 -hydroxylase activity, causing an increase in the ratio. Thus, care must be exercised when judging whether an increase in the testosterone 6 /15 -hydroxylase ratio is indicative of altered testosterone levels. We demonstrated that ketoconazole had an inhibitory effect on both testicular testosterone secretion and hepatic cytochrome P450 enzymes and that this inhibition was transient. If synthesis and inactivation of testosterone were a true balancing act, then decreased hepatic biotransformation might compensate for the decrease in gonadal testosterone synthesis, resulting in an attenuated effect on circulating testosterone levels. Results of the present study, however, demonstrated that changes in the hepatic biotransformation of testosterone had no significant effect on serum testosterone levels. Rather, changes in serum testosterone levels could be explained by changes in

IMPACT OF METABOLISM ON SERUM ANDROGEN 135 of these metabolic conversions contribute to the inactivation of the hormone. The present study demonstrated that the inhibition of certain biotransformation enzymes did not significantly impact changes in serum testosterone levels. The redundancy in biotransformation processes may be responsible for ensuring that endogenous ligands such as testosterone are efficiently cleared despite the selective inhibition of some biotransformation enzymes. Exposure to some xenobiotics can result in increased activity of some hepatic biotransformation enzymes, leading to increased clearance of testosterone (Wilson and LeBlanc, 1998). Again, this effect on biotransformation and elimination did not FIG. 6. Effects of 80 or 160 mg/kg ketoconazole treatment on serum testosterone and testicular secretion of testosterone. (A) Serum testosterone levels 4 h after treatment with either 80 or 160 mg/kg ketoconazole. (B) Rate of testosterone secreted into the incubation medium by right testes after 4 h. (C) Hepatic testosterone hydroxylation 4 h after treatment. Data are presented as mean and standard deviation (n 9 12) for each treatment level. An asterisk indicates a statistically significant difference compared to control (p 0.05). the rate of gonadal testosterone secretion. These results are consistent with our previous study demonstrating that endosulfan treatment resulted in a significant increase in hepatic biotransformation and elimination of testosterone, but had no significant effect on serum testosterone levels (Wilson and LeBlanc, 1998). A high degree of redundancy exists in the hepatic biotransformation of endogenous ligands such as testosterone. Testosterone is susceptible to hydroxylation by multiple P450 enzymes, oxidoreduction to androstenedione and androstanediols, and conjugation to either glucuronic acid or sulfate. All FIG. 7. (A) Correlation between total serum testosterone levels and testicular testosterone secretion among mice treated with 0, 80, or 160 mg/kg ketoconazole. (B) Comparison between total serum testosterone levels and hepatic testosterone hydroxylation in the same ketoconazole treated mice. (C) Effect of hepatic hydroxylation of testosterone on the relationship between total serum testosterone levels and gonadal secretion of testosterone. Data are represented as the percent of the average obtained in respective control animals.

136 WILSON AND LEBLANC impact serum testosterone levels. Gonadal testosterone secretion is regulated through serum testosterone levels by feedback control processes (Wilson and Foster, 1985). Although increased clearance of testosterone due to induction of biotransformation elimination processes may result in an initial decrease in serum testosterone levels, this decline would be detected and compensated for by an increase in gonadal testosterone secretion. The modulation of hepatic testosterone biotransformation processes would likely impact serum testosterone levels only if the feedback control of testosterone synthesis was also adversely affected. For example, chronic exposure to polychlorinated biphenyls was shown to increase hepatic 6 -hydroxylation of testosterone and to lower serum testosterone levels (Machala et al., 1998). However, this chemical treatment also inhibited gonadal CYP11A activity, the rate-limiting enzyme of steroidogenesis. Thus, the reduced serum testosterone levels were either due exclusively to the inhibition of testosterone synthesis or the combination of reduced synthesis and increased clearance. In conclusion, the results from this study indicate that testosterone levels are controlled primarily by the rate of synthesis. Chemically induced effects on hepatic steroid hormone biotransformation processes do not necessarily imply altered steroid hormone homeostasis. As noted in the introduction, there are many instances in the scientific literature where xenobiotic-mediated changes in the rate of steroid hormone biotransformation and inactivation are presumed to lead to alterations in circulating hormone levels. Our results indicate that this assertion should not be made without careful investigation. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical support provided by Cynthia V. Rider. This work was supported by the U.S. Air Force Office of Scientific Research grant No. F49620 94 1 0266 to G. A. LeBlanc. REFERENCES Baldwin, W. S., and LeBlanc, G. A. (1992). The anti-carcinogenic plant compound indole-3-carbinol differentially modulates P450-mediated steroid hydroxylase activities in mice. Chem-Biol. Interact. 83, 155 169. Bammel, A., van der Mee, K., Ohnhaus, E. E., and Kirch, W. (1992). Divergent effects of different enzyme-inducing agents on endogenous and exogenous testosterone. Eur. J. Clin. Pharmacol. 42, 641 644. Bloomquist, C. H., Kotts, C. E., and Hakanson, E. Y. (1977). Microsomal 17 -hydroxysteroid dehydrogenase of guinea pig liver: Detergent solubilization and a comparison of kinetic and stability properties of bound and solubilized forms. J. Steroid Biochem. 8, 193 198. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein dye binding. Anal. Biochem. 72, 248 254. den Besten, P. J., Elenbass, J. M. L., Maas, J. R., Dieleman, S. J., Herwig, H. J., and Voogt, P. A. (1991). Effects of cadmium and polychlorinated biphenyls (Clophen A50) on steroid metabolism and cytochrome P-450 monooxygenase system in the sea star Asterias rubens L. Aquat. Toxicol. 20, 95 110. Gagnon, M. M., Dodson, J. J., and Hodson, P. V. (1994). Ability of BKME (bleached kraft mill effluent) exposed whit suckers (Catostomus commersoni) to synthesize steroid hormones. Comp. Biochem. Physiol. 107C, no. 2, 265 273. Gonzalez, F. J. (1989). The molecular biology of cytochrome P-450s. Pharmaceut. Rev. 40, 243 288. Gradowska-Olszewska, I., Brzezinski, J. and Rusiecki, W. (1984). Excretion and peripheral metabolism of 1,2 3H-testosterone and androgens in rats following intoxication with organophosphorous insecticides: acute exposure. J. Appl. Toxicol. 4, 261 264. Grumbach, M. M., and Conte, F. A. (1981). Disorders of sexual differentiation. In Textbook of Endocrinology, 7th ed. ( J. D. Wilson and D. W. Foster, Eds.). W.B. Saunders Co., Philadelphia, PA. Guengerich, F. P., and Shimada, T. (1991). Oxidation of toxic and carcinogenic chemicals by human cytochrome P-450 enzymes. Chem. Res. Toxicol. 4, 391 407. Haake, J., Kelley, M., Keys, B., and Safe, S. (1987). The effects of organochlorine pesticides as inducers of testosterone and benzo[a]pyrene hydroxylases. Gen. Pharmacol. 18, 165 169. Hadley, M. E. (1992). Endocrinology. Simon & Schuster, Upper Saddle River, NJ. Heckman, W. R., Kane, B. R., Pakyz, R. E., and Consentino, J. (1992). The effect of ketoconazole on endocrine and reproductive parameters in male mice and rats. J. Androl. 13, 191 198. Jiritano, L., Bortolotti, A., and Bonati, M. (1986). The effect of chronic oral ketoconazole on in vivo drug metabolism in the rat. Res. Commun. Chem. Pathol. Pharmacol. 54, 173 180. Kulkarni, A. P. and Hodgson, E. (1984). The metabolism of insecticides: the role of monooxygenase enzymes. Ann. Rev. Pharmacol. Toxicol. 24, 19 42. LeBlanc, G. A. and Waxman, D. J. (1988). Feminization of rat hepatic P450 expression by cisplatin. J. Biol. Chem. 263, 15732 15739. Machala, M., Neca, J., Drabek, P., Ulrich, R., Sabatova, V., Nezveda, K., Raszyk, J., and Gajduskova, V. (1998). Effects of chronic exposure to PCBs on cytochrome P450 systems and steroidogenesis in liver and testis of bulls (Bos taurus). Comp. Biochem. Physiol. A Mol. Integr. Physiol. 120, 65 70. McMaster, M. E., van der Kraak, G. J., Portt, C. B., Munkittrick, K. R., Sibley, P. K., Smith, I. R., and Dixon, D. G. (1991). Changes in hepatic mixedfunction oxygenase (MFO) activity, plasma steroid levels and age at maturity of a white sucker (Catostomus commersoni) population exposed to bleached kraft pulp mill effluent. Aquat. Toxicol. 21, 199 218. Morita, K., Ono, T., and Shimakawa, H. (1988). Inhibitory effects of ketoconazole and miconazole on cytochrome P-450 mediated oxidative metabolism of testosterone and xenobiotics in mouse hepatic microsomes comparative study with cimetidine. J. Pharmacobiodyn. 11, 106 114. Niemegeers, C. J. E., Levron, J. C., Awouters, F., and Janssen, P. A. J. (1981). Inhibition and induction of microsomal enzymes in the rat. A comparative study of four antimycotics: miconazole, econazole, clotrimazole and ketoconazole. Arch. Int. Pharmacodyn. Ther. 251, 26 38. Pont, A., Williams, P. L., Azhar, S., Reitz, R. E., Bochra, C., Smith, E. R., and Stevens, D. A. (1982). Ketoconazole blocks testosterone synthesis. Arch. Intern. Med. 142, 2137 2140. Rodrigues, A. D., Waddell, P. R., Ah-Sing, E., Morris, B. A., Wolf, C. R., and Ioannides, C. (1988). Induction of the rat hepatic microsomal mixed-function oxidases by 3 imidazole-containing antifungal agents: selectivity for the cytochrome P-450IIB and P-450III families of cytochrome P450. Toxicology 50, 283 301. Ronis, M. J. J., Ingelman-Sundberg, M., and Badger, T. M. (1994). Induction, suppression and inhibition of multiple hepatic cytochrome P450 isozymes in the male rat and bobwhite quail (Colinus Virginianus) by ergosterol biosynthesis inhibiting fungicides (EBIFs). Biochem. Pharmacol. 48, 1953 1965.

IMPACT OF METABOLISM ON SERUM ANDROGEN 137 Sanderson, J. T., Janz, D. M., Bellward, G. D., and Giesy, J. P. (1997). Effects of embryonic and adult exposure to 2,3,7,8-tetrchlorodibenzo-p-dioxin on hepatic microsomal testosterone hydroxylase activities in Great Blue Herons (Ardea Herodias). Environ. Toxicol. Chem. 16, 1304 1310. Santen, R. J., Van den Bossche, H., Symoens, J., Brugmans, J., and DeCoster, R. (1983). Site of action of low dose ketoconazole on androgen biosynthesis in men. J. Clin. Endocrinol. Metab. 57, 732 736. Sivarajah, K., Franklin, C. S., and Williams, W. P. (1978). The effects of polychlorinated biphenyls on plasma steroid levels and hepatic microsomal enzymes in fish. J. Fish Biol. 13, 401 409. Tedford, C. E., Ruperto, V. B., and Barnett, A. (1991). Characterization of rat liver glucuronosyltransferase that glucuronidates the selective D1 antagonist, SCH23390, and other benzazepines. Drug Metab. Dispos. 19, 1152 1159. Thomson, R. G., Rawlins, M. D., James, O. F. W., Wood, P., and Williams, F. M. (1988). The acute and subchronic effects of ketoconazole on hepatic microsomal monooxygenase in the rat. Biochem. Pharmacol. 37, 3975 3980. Tredger, J. M., Smith, H. M., and Williams, R. (1984). Effects of ethanol and enzyme-inducing agents on the monooxygenation of testosterone and xenobiotic in rat liver microsomes. J. Pharmacol. Exp. Ther. 229, 292 298. van der Hoeven, T. A., and Coon, M. J. (1974). Preparation and properties of partially purified cytochrome P-450 and NADPH-cytochrome P-450 reductase from rabbit liver microsomes. J. Biol. Chem. 249, 6302 6310. vom Saal, F. S., Montano, M. M., and Wang, M. H. (1992). Sexual differentiation in mammals. In Chemically-Induced Alterations in Sexual and Functional Development: The Wildlife, Human Connection. (T. Colborn and C. Clement, Eds.), pp. 17 84. Princeton Scientific Publishing, Princeton, NJ. Waxman, D. J. (1988). Interactions of hepatic cytochromes P-450 with steroid hormones: Regioselectivity and stereoselectivity of steroid metabolism and hormonal regulation of rat P-450 enzyme expression. Biochem. Pharmacol. 37, 71 84. Waxman, D. J., Atisano, C., Guengerich, F. P., and Lapenson, D. (1988). Human liver microsomal steroid metabolism: identification of the major microsomal steroid hormone 6b-hydroxylase cytochrome P-450 enzyme. Arch. Biochem. Biophys. 263, 424 436. Waxman, D. J., Ko, A., and Walsh, C. (1983). Regioselectivity and stereoselectivity of androgen hydroxylation catalyzed by cytochrome P-450 isozymes purified from phenobarbital-induced rat liver. J. Biol. Chem. 10, 11937 11947. Wilson, J. D., and Foster, D. W. (1985). Textbook of Endocrinology. W.B. Saunders Company, Philidelphia. Wilson, V. S., and LeBlanc, G. A. (1998). Endosulfan elevates testosterone biotransformation and clearance in CD-1 mice. Toxicol. Appl. Pharmacol. 148, 158 168. Wilson, V. S., McLachlan, J. B., Falls, J. G., and LeBlanc, G. A. (1999). Alteration in sexually dimorphic testosterone biotransformation profiles as a biomarker of chemically induced androgen disruption. Environ. Health Perspect. 107, 377 384.